Experts are increasingly turning to machine learning to predict antibiotic resistance in pathogens. With its help, resistance mechanisms can be identified based on a pathogen’s genetics. However, the results should be viewed with caution: Researchers at the Helmholtz Institute for RNA-based Infection Research (HIRI) in Würzburg have shown that the models are often less reliable than assumed. Their findings were published in the journal PLOS Biology. They contribute to the development of more reliable tools for predicting and combating antibiotic resistance.

At first glance, biology and computer science seem like opposites. But wherever enormous amounts of data are generated from research, progress is hardly possible without digital methods. Bioinformatician Prof. Andreas Keller therefore relies on artificial intelligence. He heads the department “Clinical Bioinformatics” at the Helmholtz Institute for Pharmaceutical Research Saarland (HIPS). In this episode of InFact, he talks about how AI can help us understand how beneficial and harmful bacteria communicate with each other in our bodies, how to predict when infections will cause long-term effects, and how this can be used to develop new drugs against dangerous pathogens.

Biphenomycins, natural products derived from bacteria, show excellent antimicrobial activity, but have long remained out of reach for drug development. The main obstacle was the limited understanding of how these compounds are produced by their microbial hosts. A research team led by Tobias Gulder, department head at the Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), has now deciphered the biosynthetic pathway of the biphenomycins, establishing the foundation for their pharmaceutical advancement. The team published its findings in the journal Angewandte Chemie International Edition.